stbenjam OOXML Validator
Unpack the document, then validate every part against the OOXML schemas:
unzip -q report.docx -d unpacked/
python scripts/validate.py unpacked/ --report validation.html
Reading the output
Each failure names the part, the line, and the schema rule it violated.
Fix the part in unpacked/, re-zip, and re-run until the run is clean.
Pass --strict to fail on the warnings the Office applications tolerate,
such as an unknown attribute in a namespace the document declares but
never uses.
Legacy flat-file exports are out of scope: convert them to .docx first.